INDEX

Explanations

attends to the token "request" from various types of token phrases

oai_attention-head · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 0-gemmascope-att-16k

Configuration

google/gemma-scope-2b-pt-att/layer_0/width_16k/average_l0_104

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.0.attn.hook_z

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Head Attr Weights

0:0.80

1:0.01

2:0.01

3:0.02

4:0.08

5:0.01

6:0.01

7:0.02

Negative Logits

 them

-0.40

 their

-0.30

it

-0.29

for

-0.28

him

-0.28

his

-0.27

dibles

-0.27

 THEM

-0.26

 then

-0.26

Mereka

-0.26

POSITIVE LOGITS

 surla

0.42

StoryboardSegue

0.41

 nakalista

0.40

 &___

0.40

 linkovi

0.39

 wireType

0.38

audiovisuel

0.38

 <<<<<<<<<<<<<<

0.38

getDoctrine

0.38

Portale

0.37

Activations Density 0.024%

attends to the token "request" from various types of token phrases

No Comments

No Known Activations